https://github.com/kremerj/gan

A 1D toy example of optimizing a generative model using the WGAN-GP model.
https://github.com/kremerj/gan

generative-adversarial-network tensorflow-models toy-example

Last synced: 8 months ago
JSON representation

A 1D toy example of optimizing a generative model using the WGAN-GP model.

• Host: GitHub
• URL: https://github.com/kremerj/gan
• Owner: kremerj
• Created: 2017-07-20T15:14:48.000Z (almost 9 years ago)
• Default Branch: master
• Last Pushed: 2017-07-24T19:48:44.000Z (almost 9 years ago)
• Last Synced: 2025-04-07T02:21:56.197Z (about 1 year ago)
• Topics: generative-adversarial-network, tensorflow-models, toy-example
• Language: Python
• Size: 3.28 MB
• Stars: 23
• Watchers: 1
• Forks: 11
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# 1D Generative Adversarial Network Example

![Example output of the WGAN-GP optimization](animation.gif)

Generates a GAN to approximate a 1D Gaussian using the

improved [Wasserstein GAN (WGAN-GP)](https://arxiv.org/abs/1704.00028). Based on the codes from the

repositories of [Ishaan Gulrajani](https://github.com/igul222/improved_wgan_training)

and [AYLIAN](https://github.com/AYLIEN/gan-intro) and the blog posts by

[Eric Jang](http://blog.evjang.com/2016/06/generative-adversarial-nets-in.html) and

[John Glover](http://blog.aylien.com/introduction-generative-adversarial-networks-code-tensorflow).

## Running the example

This code uses Python 3.6.1. Install the dependencies via

    $ pip install -r requirements.txt

To generate the animated gif, you need to install the ImageMagick library.

To run the code just run

    $ python main.py

and it should produce output similar to the animation above.